When deciding on the extent and means of pursuing avenues for improving welfare outcomes, the social planner must take into consideration constraints in the form of biophysical relationships and behavioral, institutional and political constraints. The biophysical relationships consist of several elements. First is the production func- tion, which links outputs to ecological inputs and weather. One of the key chal- lenges in designing agricultural policies is in understanding the heterogeneous impacts of climate change on productivity. Furthermore, modeling of the produc- tion function needs to consider both continuous as well as discrete variables. This approach allows us to investigate technology adoption in response to climate change (Mendelsohn and Dinar 1999; Antle and Capalbo 2010; Arslan et al. 2015).
Understanding the stochastic nature of the production function, particularly due to weather realizations, will also be important in designing programs, such as insur- ance and inventory, to address the challenges of climate change. The second bio- physical element is the externality function, which expresses the relationships between economic activities and the various externalities generated by them
2 We basically adopt the IPCC WGII AR5 definitions of vulnerability and resilience, as provided in Appendix 1. However, for conceptual convenience, we are defining vulnerability as a household- level characteristic, and resilience as a system-level characteristic.
(Zilberman 2014). In the context of CSA, the greenhouse gas emissions are the main, but not sole, externality considered. Various agricultural practices and invest- ments also generate both positive and negative local externalities. Overuse of inor- ganic fertilizer generates greenhouse gas emissions and can also pollute local water sources (Norse 2012). Investment in soil and water conservation structures at the farm and ecosystem levels can generate positive spillover benefits to neighboring farmland productivity (Mirzabaev et al. 2015; McCarthy et al. 2012). Without effec- tive coordination and collective action, too few positive spillovers, and too many negative spillovers, will be generated.
In analyzing both the production and externality functions, we recognize that agriculture is very diverse, and different sectors of agriculture (e.g. irrigated agricul- ture, rain-fed agriculture, etc.) will experience climate change differently. Livestock husbandry and fisheries will have unique challenges as well, and our analysis should strive to provide appropriate solutions that recognize specific contexts.
The behavioral constraints include market choices made by risk-averse individ- ual agents (both inputs and outputs) operating in contexts where insurance markets are very thin or entirely absent. Our analysis will emphasize the importance of cli- mate conditions on the supply and demand of various goods. The choices will be dependent on risk preferences and market conditions, as well as government poli- cies. An important category of behavioral choices relates to decisions regarding technology adoption, including irrigation, seed varieties and production practices.
Almost all empirical evidence suggests that uninsured risk and uncertainty leads to low levels of adoption of new technologies, and this behavioral constraint must be addressed if hoped-for wide-scale adoption is to be realized (Antle and Crissman 1990; Dercon and Christiansen 2011). Furthermore, adopting any new technology is often itself seen to be risky by the farmer who faces uncertainty about its perfor- mance (Foster and Rosenzweig 2010). Zilberman et al. (2012) note that, in addition to risk preferences, the diffusion of technology adoption as an adaptation to climate change will also be a function of heterogeneity in farmers’ access to capital, the underlying agro-ecology, and prevailing institutions that can foster or hinder adoption.
Technology adoption and institutional innovations are also a function of political constraints. As Hayami and Ruttan (1971) emphasize, innovations of new technolo- gies are outcomes of economic choices that are responsive to incentives and poli- cies. Thus, the literature on innovation also emphasizes the role of learning in innovation and the evolution of new technologies, which in turn affect adoption.
Political economic modeling suggests that government policy is affected by eco- nomic conditions as well as environmental and political considerations (Buchanan and Tollison 1984; Shepsle 1992; Rausser et al. 2011). These suggest that individual government policy choice problems are derived from their own political economy constraints so that the decision to implement policies that favor certain technologies over others will be a function of this political calculus. Where political weighting favors high economic growth, for instance, the technologies promoted may conflict both with resilience and low-emissions growth goals, for instance.
In addition to political economy considerations, additional political constraints will bound the range of feasible policy and legal actions to address climate change.
Some policy solutions to climate change may not be politically feasible, and realis- tic policy design must consider feasibility of solutions within various local and global contexts. For example, it will be politically easier and it makes common sense to enact policies that improve human well-being and welfare regardless of climate change. A no-regret constraint may bind the set of policies that would be valuable under certain future conditions to those that also address pressing issues of food security or sustainable land use, thereby satisfying distributional and environ- mental objectives.
The institutional constraints include input, output and labor markets, property rights and tenure security, information dissemination systems such as agriculture extension and weather forecasting, credit and insurance markets and their regula- tory framework, social safety net programs, environmental regulations, and the international trading system and local import, export, and foreign direct investment regulations. The institutional environment has a significant impact on farmers’
incentives and ability to invest in agriculture practices with CSA characteristics and to adapt to climate change. Thin value supply chains limit farmers’ ability to access inputs in timely fashion, and sell their output at a profit. Integrated supply chains can significantly reduce market price swings in response to extreme weather events, thereby reducing vulnerability of rural households to poor crop output and high food prices (Reardon and Timmer 2007). As discussed above, thin or absent credit markets, often combined with very limited insurance mechanisms, dampen incentives to make any types of investment on-farm, and limits the choices avail- able to risk-averse farmers to adapt. Similarly, property rights systems that result in tenure insecurity also limit incentives to invest in land (Mirzabaev et al. 2015;
Holden et al. 2009).
The ability to adapt to climate change will also be affected by the information dissemination system and farmers’ ability to access weather forecasts and longer- term climate predictions and to incorporate that information into adaptation and coping strategies. Additionally, improving the resilience of the agricultural system as a whole will necessitate making investments and coordinating changing practices at scales higher than the household level. The ability to invest in larger-scale infra- structure to improve the resilience of a watershed (Bassist et al. forthcoming), or coordinating investments in tree planting or check dams across many small com- munities will depend on local property rights, land use regulations and powers of eminent domain, as well as environmental regulations. The ability to coordinate actions across communities will also be affected by collective active institutions and local-level governance structures (Meinzen-Dick et al. 2004; Pender et al. 2006).
The ability to relax institutional constraints will be key in reducing household vul- nerability and increasing system resilience in many contexts.
The optimization problem has several dynamic constraints as well. The first con- straint is the dynamics of climate change. Because of the nature of agriculture, it is important to have an adequate assessment of climatic variation over space and time in
order to make predictions of yields and outputs. There is much uncertainty in climate modeling and it must be incorporated into policy design. Thus, it is not sufficient to get average predictions of climatic patterns over time, but also some indication of variability and reliability thereof. Uncertainty of weather patterns is important because as Dixit and Pindyck (2001) suggested, the pattern and levels of uncertainty delay the optimal timing of investment. With uncertainty, decision- makers value addi- tional information and are willing to wait some time for more information, which can lead to significant delays in investments. This compounds risk-averse farmers’ disin- centives to invest in land or adopt new technologies.
A second dynamic element is population growth, which affects demand for food as well as urbanization patterns, both of which are important determinants of opti- mal agricultural growth pathways. Human population growth is also behavioral to some extent and thus population dynamics must take account of behavioral param- eters. Furthermore, population dynamics are subject to uncertainty so we must con- sider outcomes under several scenarios in assessing and designing climate change policies.
The third dynamic element is the ongoing transition in agriculture associated with globalization and the spread of information and technological advances. Global supply chains are spread everywhere, and the expanded use of the internet, cell phones, and improved transportation mechanisms are likely to continue.
Technological change is especially important given the role of innovation and adop- tion in adaptation to climate change, but its diffusion will be a function of both political constraints as well as the need to adapt technologies to site-specific charac- teristics. One also needs to understand the workings of the supply chain innovations in different regions and how they can be utilized to introduce new technologies in response to climate change. While further integration and connectivity can increase agricultural system resilience by reducing, pooling and transferring risks, positive results will nonetheless be a function of the international and national level regula- tory frameworks. To achieve food security objectives, such frameworks need to incorporate regulations that limit monopolistic/oligopolistic power and instead har- ness the risk-reducing benefits for everyone in the agricultural system, as well as effective enforcement mechanisms.